There may have been some other early ones; see e.g. IEN-3 Jon Postel "Meeting Notes 15 August 1977". I was thinking of the TCP's which showed up for the first TCP Bakeoff, on 27 January, 1979. 1982? That's almost up to Linux time, in early internet time! :-) It's like 'dog years'; a year back then was an aeon! OK, to start with, a few 'definitions': TCP 2 - the earliest version that was really implemented much; no separate IP layer TCP 2.5 - basically the same as TCP 2 (packet formats, algorithms), but with things re-labelled into 'TCP' and 'IP' TCP 3 - variable length addresses (later discarded, sadly) (The archaeology in the IENs is a little complicated, because meeting notes aren't in temporal order, IEN-number-wise. E.g. the notes for the meeting on 2-4 August 1978 (the first one I went to, BTW) are IEN-53, which came out on 21-Aug-78; but the notes for the 14-15 July 1977 meeting are IEN-65, which came out on 5-Aug-78.) Anyway, about the first details about the TENEX TCP implementation (done by Bill Pluumer, whom we have now sadly lost) I could find are in IEN-67 ("Meeting Notes - 30 & 31 January 1978"): https://www.rfc-editor.org/ien/ien67.pdf where he reports that an implementation of TCP-2 (in the OS, not in a user process, as an early version was). In IEN-53 ("Meeting Notes - 2,3&4 August 1978"): https://www.rfc-editor.org/ien/ien53.pdf Action Items 6 and 7 refer to 6) the installation of presumably somewhat working TCP 2.5's at SRI, BBN and SRI, and 7) the installation of test version of TCP 4 at BBN. Finally, IEN-57: https://www.rfc-editor.org/ien/ien57.pdf by Bill Plummer, "Provisional TCP Development Plan" (for TENEX/TOPS-20 TCP) gives sime interesting details. Noel